Contents lists available at ScienceDirect

Cognitive Development

journal homepage: www.elsevier.com/locate/cogdev

Who do I believe? Children’s epistemic trust in internet, teacher,and peer informantsFuxing Wanga,⁎, Yu Tonga, Judith Danovitchb

a School of Psychology, Central China Normal University, Wuhan, Hubei, 430079, ChinabDepartment of Psychological and Brain Sciences, University of Louisville, Louisville, KY, 40292, USA

A R T I C L E I N F O

Keywords:Epistemic trustInformantsInternetTeacherPeer

A B S T R A C T

In the cultural context of rapidly increasing internet access, two experiments examine how 5- to8-year-old Chinese children and adults evaluate information from an unspecified internet sourceor a known human informant (a teacher or a peer). In Experiment 1, when evaluating statementsfrom a variety of domains, adults regarded the internet and a teacher as more trustworthy than apeer. Younger children did not show differential endorsement of statements by any source, andolder children endorsed statements attributed to a teacher over those from the internet. InExperiment 2, when the statements involved scientific and historical facts only, all age groupssought out and endorsed information from the internet or a teacher more often than from a peer,but only adults trusted the internet more than the teacher. These results demonstrate that chil-dren can reason about the reliability of informants across different categories and that their trustin information sources is contingent on the type of information being presented.

1. Introduction

Information is indispensable for human survival and development, and this is especially true for children. In order to survive andadapt to complicated social and natural environments, children can obtain information from direct observations and their personalexperiences. However, due to practical limitations, children also have to gain massive amounts of information via other sources (seeBrosseau-Liard, 2017; Harris, 2012). These sources include other people, such as parents, teachers, and peers, and media sources,such as books, TV, and the internet. In some cases, different information sources provide nearly identical answers to the samequestion (e.g., Giles, 2005). However, in other cases, different information sources provide inconsistent answers (Enesco, Rodríguez,Lago, Dopico, & Escudero, 2016). Being able to select and evaluate information is a core skill for people in the 21 st century (van Laar,van Deursen, van Dijk, & de Haan, 2017), and this challenge is compounded by the proliferation of relatively new and unfamiliarinformation sources such as the internet. In the current studies, we examine how the ability to choose between a known individual(i.e., a teacher or a peer) and an unspecified internet source develops, and the extent to which children and adults trust an unspecifiedinternet source when the information it provides conflicts with what a known individual says.

The rapid growth of the internet has facilitated the creation and spread of information (Johnson, 2007), yet it also poses newchallenges. Because information on the internet can be authored anonymously and easily published or modified, the availability offabricated or false quotations, images, and other types of counterfeit information on the internet has increased tremendously(Flanagin & Metzger, 2000; Mehrabi, Hassan, & Ali, 2009; Metzger & Flanagin, 2013; Waldrop, 2017). Due to the growing popularityof portable electronic devices (e.g., tablets, smartphones), there are also more opportunities for children to be exposed to and to

https://doi.org/10.1016/j.cogdev.2019.05.006Received 20 July 2018; Received in revised form 13 May 2019; Accepted 13 May 2019

⁎ Corresponding author at: 152 Luoyu Street, Tianjiabing Bldg 302, Central China Normal University, Wuhan, Hubei, 430079, PR China.E-mail address: fxwang@mail.ccnu.edu.cn (F. Wang).

Cognitive Development 50 (2019) 248–260

Available online 22 May 20190885-2014/ © 2019 Elsevier Inc. All rights reserved.

T

access the internet. Indeed, middle- and high-school students have become increasingly reliant on the internet as a means of obtaininginformation (St. Jean, Taylor, Kodama, & Subramaniam, 2017).

In China, children’s access to the internet has dramatically increased over the past 10 years. In 2017, 91% of 3- to 6-year-oldChinese children had accessed the internet through smartphones, tablets and computers, and more than 60% of 5-year-old Chinesechildren used the internet for entertainment for at least 30min every weekend (China Internet Network Information Center, 2018b).Constrained by their literacy skills, Chinese children mainly access the internet through online video platforms and have limitedexperience looking up information online until at least age 9 (China National Youth Palace Association, 2015). Chinese children’sonline activities are not as diverse as adults’ (e.g., children seldom use social media or do online shopping), yet they still haverelatively unlimited access to popular information on the internet on websites other than those designed expressly for children.Although information is strictly censored by the China Internet Network Information Center to make sure it contains no sexuality,violence, or other kinds of illegal content, the authenticity and the veracity of the information available to Chinese children stillvaries. Moreover, though most Chinese parents set strict limits to their children’s amount of internet use to protect children’s healthand prevent internet addiction, having internet access is still a recent innovation for most Chinese families. Chinese parents have verylimited awareness of the need to help children evaluate the reliability of information found online (More than 60% Chinese childrenhave mobile phones, 2016). Given children’s strong propensity to imitate and learn from others (Harris, 2012), some Chineseeducators are also concerned about the potential consequences of children accessing unreliable information online, especially in theabsence of formal or informal internet instruction (More than 90% of children get access to internet, 2017). Thus, in contrast to otherdeveloped countries such as the United States, where internet access has increased relatively steadily over the past few decades andguidelines have been in place for educating children about how to safely and effectively use the internet since as early as 1998(although local adoption of these guidelines has varied; International Society for Technology in Education, 1998, 2007), China is inthe midst of a transition where adults have not necessarily had extensive experience with the internet and the messages that childrenreceive about the reliability of the internet as an information source are limited.

Despite the fact that Chinese 5- to 8-year-olds have been exposed to the internet since early in their lives, their understanding andattitudes towards the internet as an information source have not been studied. Research with children in other countries suggests thatchildren in this age range do not understand how the internet works (Mertala, 2019; Yan, 2005), nor are they very effective atobtaining information online (Dodge, Husain, & Duke, 2011; Duarte Torres, Weber, & Hiemstra, 2014). Moreover, even when theyare capable of retrieving information from the internet, older children and adolescents find it difficult to evaluate the credibility ofthe information they encounter (Eastin, Yang, & Nathanson, 2006; Metzger, Flanagin, Markov, Grossman, & Bulger, 2015). Byexamining Chinese children’s beliefs about the relative credibility of information attributed to an unspecified internet source, thestudies presented here address important questions about the formation of children’s attitudes towards information sources in theabsence of education or adult input about the reliability of those sources.

Young children can select and evaluate information based on the characteristics of an informant (Bascandziev & Harris, 2016;Enesco et al., 2016; Landrum, Cloudy et al., 2015; Landrum, Eaves et al., 2015), yet little is known about how children judgeinformation provided by other people relative to technology-based sources. The current study examines children’s preferences for anunspecified internet source or a known individual, and their endorsement of information provided by an unspecified internet source,a teacher, or a peer. Previous studies have found that children as young as age 3 can weigh factors such as prior accuracy whenjudging informants (e.g., Koenig, Clément, & Harris, 2004; Koenig & Harris, 2005). By ages 4 and 5, they also take into accountadditional characteristics, including informants’ familiarity (Corriveau & Harris, 2009; Danovitch & Mills, 2014), attractiveness(Bascandziev & Harris, 2014, 2016), and benevolence (Johnston, Mills, & Landrum, 2015; Li, Heyman, Xu, & Lee, 2014; Palmquist,Jaswal, & Rutherford, 2016). In these studies, children typically encounter pairs of informants from the same category. These aremost often pairs of people, though there are a few studies using pairs of fictional characters (e.g., Danovitch & Mills, 2014) andreliable or unreliable computers (Danovitch & Alzahabi, 2013). In the current studies, we examine children’s preferences and en-dorsements in situations where the informants belong to different categories.

Given that assessing the reliability of information sources is of great importance to children’s early learning (Harris, 2012),Brosseau-Liard (2017) suggests that children’s source evaluation skills form the basis for later-developing critical thinking abilities,and that source evaluation skills should not be limited to human sources. A few pertinent studies indeed have found that children canapply the same heuristics that they use to evaluate the reliability of human informants to reason about the credibility of otherinformation mediums. For instance, preschoolers preferred information provided by a previously accurate computer rather than apreviously inaccurate one (Danovitch & Alzahabi, 2013) and trusted a book with logically consistent contents rather than a book withlogically inconsistent contents (Doebel, Rowell, & Koenig, 2016). Moreover, Einav and colleagues revealed that early readers treatedoral information supported by printed text as more reliable than oral information not supported by print (Corriveau, Einav, Robinson,& Harris, 2014; Einav, Robinson, & Fox, 2013; Eyden, Robinson, Einav, & Jaswal, 2013; Eyden, Robinson, & Einav, 2014; Robinson,Einav, & Fox, 2013). These findings suggest that it is reasonable to consider children’s source evaluation strategies as fundamentalelements of social learning that enable children to learn more reliable information (see Harris, Koenig, Corriveau, & Jaswal, 2018 fora review). If this is the case, children’s strategies might be domain-general methods which are not unique to evaluating humansources, and children should also be capable of applying them to assess the credibility of informants from different ontologicalcategories. However, little is known about whether children treat one category of information source (e.g., internet sources) as morereliable than another (e.g., known individuals).

In this study, we investigate how children reason about information provided by an unspecified internet source in contrast withinformation provided by a teacher and a peer. Young children view teachers as emotional referents (Koomen & Hoeksma, 2003;Mantzicopoulos & Neuharth-Pritchett, 2003) and, in some cases, form attachments to them (see Bergin & Bergin, 2009 for a review).

F. Wang, et al. Cognitive Development 50 (2019) 248–260

249

Moreover, children view teachers as authority figures with a willingness to impart reliable facts in the classroom (Enesco et al., 2016),and Chinese culture strongly supports this view. Because Chinese children’s experience of the internet seems to largely consist ofwatching videos or playing games (rather than interacting with others or using social media), children may view the internet as animpersonal source that does not have the pedagogical capacity or authority of a teacher. If this is the case, children should showgreater trust in information from a teacher than from the internet. On the other hand, if children prioritize information capacity (suchas the abundance of information and the ease of accessing information, see She, Liu, & Fan, 2018 for a review) over pedagogicalauthority and social bonds, they should show greater trust in the internet than a teacher. This outcome would suggest that episte-mological characteristics are most important when children evaluate the reliability of an unspecified internet source in comparison toa known individual. It is also possible that children will treat a teacher or the internet as equally credible, although they may trusteach source for different reasons.

In order to further disambiguate the role of knowledge access from that of a social relationship, we also examine children’s trust inan unspecified internet source or a teacher when faced with information from a peer. Given that children generally treat adults asmore knowledgeable than other children (Jaswal & Neely, 2006), if children show more trust in a same-age peer than in the internet,it would suggest that social connection is the crucial factor in determining children’s trust (Elashi & Mills, 2014), but if children treatsocial bonds with peers as less important than access to extensive amounts of information, they should be more likely to believe ininformation from the internet. Our design also includes trials where participants choose between a teacher and a peer in order toascertain that children in our sample view teachers as more credible than peers.

Participants in the current studies include 5- to 6-year-old kindergarteners and 7- to 8-year-old second-graders. We focused onthese age groups because older children would presumably have had more exposure to the internet and be more capable of accessingthe internet on their own. There are indeed some studies suggesting that children become more capable of evaluating informationfrom the internet as they grow older (Eastin et al., 2006; Flanagin & Metzger, 2000; Hart & Liu, 2003; Metzger et al., 2015) and thatimprovement in these skills is already apparent between kindergarten and 2nd grade (Dodge et al., 2011). Moreover, older childrenare more likely to express concerns about the accuracy of online information than younger children (Yan, 2006). Thus, we wereinterested in exploring the relation between age and trust in an unspecified internet source or a known individual source. In addition,we examine whether children’s online experiences related to their epistemic trust in information coming from an unspecified internetsource.

2. Experiment 1

2.1. Participants

Thirty 5- to 6-year-old kindergarteners (M age = 5;11, range=5;7–6;2, 15 girls) and 30 7- to 8-year-old second-graders (M age =8;0, range= 7;0–8;9, 19 girls) participated. Five additional kindergarteners took part but were not included in the data analysisbecause they refused to endorse either informant and insisted on their personal answers to the questions instead. Most of the childrencame from middle-class families in the city of Wuhan, China. A priori power analysis using G*Power 3 (Faul, Erdfelder, Lang, &Buchner, 2007) revealed that a minimum of 54 total child participants would be needed to detect a medium effect size of 0.25 (avalue based on a prior study showing that older children were more likely to choose a reliable computer than younger children;Danovitch & Alzahabi, 2013; Exp.2) within a 2 (Age Group: kindergarten, second-grade) × 2 (Alternate Informant: teacher, peer)ANOVA with power (1−β) set at .95 and α set at .05. To assess what a mature pattern of responses would be for our tasks, 30 collegestudents (M=20.30 years; SD=1.90; 20 females) were also recruited from a university in central China. They represented differentlevels of seniority and various majors, including psychology, mathematics, physics, and literacy. All participants received gifts orwere paid for their participation. The study was approved by the Ethical Institutional Review Board of Central China Normal Uni-versity.

2.2. Materials

Because in most circumstances, children can use the internet to access a wide range of information and Chinese children areunlikely to have received any explicit instructions about the reliability of internet sources for specific types of information, thisexperiment included trivia from a broad range of topics. The 12 pieces of trivia that were included involved obscure or difficult toobtain answers that were not contrary to common sense (e.g., people are more likely to be bitten by mosquitoes when they are eatingbananas). Each piece of trivia was adapted to two statements that differed in one detail (e.g., “people are more likely to be bitten bymosquitoes when they are eating bananas” and “people are more likely to be bitten by mosquitoes when they are eating apples”; seeAppendix A for English translations of the Chinese statements used). For each pair of statements, there was a related photographprinted on a card (e.g., a photo of a mosquito).

2.3. Procedure

The experiments were run in a quiet lab and sessions were recorded for coding purposes.

2.3.1. Children’s online experienceAfter the participants completed a demographic questionnaire, they were asked about their online experience. First, participants

F. Wang, et al. Cognitive Development 50 (2019) 248–260

250

were asked an open ended question: “Have you used computers or iPads or iPhones? What did you do with them?” Then, they wereasked a series of follow-up questions probing their online experiences (e.g., did you watch videos on Youku or IQiyi? Did you look upinformation online?) that were based on common online activities listed in The 42nd China statistical report on Internet development(China Internet Network Information Center, 2018a).

2.3.2. Informant’s descriptionThe experimenter described each informant to participants. The internet was described as “when you use computers and other

portable devices such as iPads or iPhones to obtain information from IQiYi or Youku (a video platform where people can publishvideos, similar to YouTube) or a news website or Baidu (a search engine similar to Google that is widely used by children in China).”(All descriptions are translated from Chinese.) These examples of internet-based sources were used because our observations andexperience suggest that even if children have used the internet, they may not be familiar with the word “internet.” The teacher wasdescribed as “a teacher who taught you” and the peer was described as “a child like you.” In addition, to ascertain whether childrenhad understood the meaning of each informant, they were required to give an example of retrieving information from the internet, ateacher, and a peer (for some typical examples of children’s reports, see results section).

2.3.3. Endorse trialsAfter children reported appropriate examples of information retrieval, the experimenter explained that participants would hear

some interesting statements coming from the internet, a teacher, or a peer, and they would indicate which statement they believed.During each trial, participants were first asked whether they knew the answer to the trivia question (e.g., what color do you thinkAmericans like best?; see Appendix A). None of the participants reported answers consistent with statements that were later pre-sented. Then, the experimenter read out loud two statements which were adapted from one piece of trivia and attributed thestatements to two of the three informants (e.g., “Teacher says that most Americans like yellow best; according to the internet, mostAmericans like purple best”; translated from Chinese, see Appendix A). Participants were then asked to endorse one of the claims (e.g.,“Now, what color do you think Americans like best, yellow or purple?”; see Appendix A). In order to maintain children’s attention andhelp them to understand each piece of trivia, the experimenter presented a related photograph when the statement was introduced.

There were a total of 12 trials, divided into 4 blocks with 3 statements each (see Appendix A). In each block, one pair ofstatements were presented by the unspecified internet source and a teacher, another pair were presented by the internet source and apeer, and the remaining pair of statements were presented by a teacher and a peer. Which topics were presented by which pair ofsources was balanced across participants, yielding 6 item orders within each block. In addition, the ordering of the blocks and theordering of informant presentation in each block was counterbalanced across participants. For each participant, informant A followedinformant B for half of the blocks, and for the other half of the blocks, informant B followed informant A.

2.4. Results

2.4.1. Children’s online experienceTwo trained coders (both graduate level researchers, one of whom was blind to the research goals and manipulation) coded each

child’s online experiences into three non-exclusive categories (watching online videos, using search engines and online news sources,using social media or chatting online) based on the child’s responses to both the open-ended and forced choice questions. There was100 percent agreement between coders.

Apart from generating appropriate examples of retrieving information from a teacher (e.g., “Last class, I learned from my teacherthat horses usually sleep standing up”) and a peer (e.g., “I asked my friend how to fold a paper airplane”), all children gave ap-propriate examples of using the internet (e.g., “I used it to watch short video on iPad about how various flowers went to seed orsomething unusual”), including watching videos and looking up information online. When asked to report their personal experienceusing the internet, 29 kindergarteners (1 child’s recording could not be heard clearly) indicated watching online videos on Youku,IQiyi, or other video platforms. As for second-graders’ online experience, 14 seconds-graders only reported watching online videos,while 13 seconds-graders reported watching videos and looking up information online (three additional second-graders’ recordingswere damaged or lost). Notably, no child in either age group mentioned using WeChat, a popular Chinese multi-purpose messagingand social media platform, or other social media.

To verify whether children’s reports were reliable, children’s parents were also contacted via phone or e-mail. Parents indicatedthat when online, their children mainly watched short videos or stories on video platforms, with no personal experience of any socialmedia. Teachers in the elementary school and kindergarten where children were recruited were also interviewed. They indicated thatthey did not provide any formal education about the internet and that using the internet to retrieve information was not emphasizedin classrooms. Moreover, although internet access was available in the classroom, students were not allowed to take any portableelectronic devices into school.

2.4.2. Children’sendorsement of information from the internetIn order to see whether children showed stronger endorsement of information from the unspecified internet source when it was

paired with a teacher or a peer, we computed the number of times participants chose to endorse the statement attributed to theinternet in the internet vs. teacher or internet vs. peer trials (max=4 for each set of trials). Preliminary analyses did not show aneffect of gender (Kindergarteners: ts< 1.57, ps> 0.12; Second-graders: ts< 1.77, ps> 0.09) nor item order (Kindergarteners:Fs< 2.51, ps> 0.09; Second-graders: Fs< 2.56, ps> 0.06) on children’s endorsements, so these variables were not included in

F. Wang, et al. Cognitive Development 50 (2019) 248–260

251

subsequent analyses. To investigate whether the alternate informant (teacher or peer) impacted children’s endorsement of statementsfrom the internet, we conducted a 2 (Age Group: kindergarten, second-grade) × 2 (Alternate Informant: teacher, peer) ANOVA.Neither the main effect of age (F(1, 58)= 3.00, p= .089, ηp2 =.049) nor alternate informant (F(1, 58)= 0.95, p= .333, ηp2 = .016)was significant. However, the age by alternate informant interaction was significant, F(1, 58)= 8.59, p= .005, ηp2 = .129. Furtheranalysis of the interaction using simple effects tests indicated that second-graders were more likely to endorse statements from theinternet when the alternate informant was a peer than when it was a teacher (F(1, 58)= 7.63, p= .008).

In the internet vs. teacher trials, second-graders endorsed statements from the internet at rates lower than chance (chance= 2, t(29)= 3.17, p= .004, d=0.58; see Table 1). However, kindergarteners’ responses were no different from chance, t(29) = −1.27,p= .214, d = -0.23, suggesting similar levels of trust in each informant. In the internet vs. peer trials, both second-graders’ andkindergarteners’ responses did not differ from chance (Second-graders: t(29)< 1, p > .05, d=0.08; Kindergarteners: t(29)< 1,p > .05, d=0.09; see Table 1). This result reveals that kindergarteners and second-graders did not show differential trust in internetand peer sources.

In order to examine whether children’s online experience related to their endorsement of statements from the internet, wecompared scores between second graders who reported having looked information up online and those who did not. The resultssuggested that children’s experience with the internet did not relate to their endorsements of statements from the internet (see TableS1 in Supplementary Materials).

2.4.3. Children’s endorsement of information from a teacher versus peerFor the four trials where statements were attributed to the teacher or the peer, responses were assigned 1 point each time the child

endorsed the teacher’s response, yielding a score of 0–4. Both second-graders and kindergarteners were at chance in choosingbetween the teacher’s and peer’s statements (Second-graders: t(29)= 0.28, p= .778, d=0.05; Kindergarteners: t(29)= 1.78,p= .086, d=0.33). There was also no difference between kindergarteners and second-graders’ rate of endorsing the teacher’sstatements, t(58)= 0.96, p= .341, d=0.24.

2.4.4. Adult endorsementsConsidering the major differences in online experience between adults and children, we analyzed the data about adults’ choices

separately to explore what a mature pattern of responses looks like. Preliminary analyses did not show an effect of gender (ts< 1.04,ps> 0.31) nor item order (Fs< 2.48, ps> 0.06), so these variables were not included in subsequent analyses. We found that collegestudents considered the unspecified internet source and a teacher as equally reliable in the internet vs. teacher trials, t(29)= 1.19,p= .243, d=0.21, but they were less likely than chance to endorse the statements attributed to the peer in both the internet vs. peertrials and the teacher vs. peer trials (chance=2, internet vs. peer: t(29)= 3.43, p= .002, d=0.63; teacher vs. peer: t(29)= 5.14,p < .001, d=0.94).

Table 1Endorsement of Informants in Each Set of Trials (maximum=4, chance= 2).

Age Informant Endorsement

Internet (vs. Teacher) Internet (vs. Peer) Teacher (vs. Peer)

M SD M SD M SD

Kindergarten 2.23 1.01 1.90 1.09 2.37 1.23Second-grade 1.40* 1.04 2.07 0.83 2.07 1.29Adult 1.77 1.07 2.70* 1.12 2.90* 0.96

* indicates a significant preference (p < .05) for one informant.

Table 2Frequencies of Choices in Ask and Endorse Items in Experiment 2 (maximum=4, chance=2).

Question Type Age Informant Choice

Internet (vs. Teacher) Internet (vs. Peer) Teacher (vs. Peer)

M SD M SD M SD

Ask Trials Kindergarten 1.97 1.49 3.32* 1.25 3.68* 0.60Second-grade 1.83 1.21 3.77* 0.50 3.80* 0.55Adult 3.53* 0.97 3.83* 0.38 2.83* 1.18

Endorse Trials Kindergarten 1.71 1.24 2.84* 1.16 3.42* 0.85Second-grade 1.63 1.35 3.60* 0.77 3.63* 0.67Adult 2.70* 1.29 3.47* 0.94 3.27* 0.94

* indicates a significant preference (p < .05) for one informant.

F. Wang, et al. Cognitive Development 50 (2019) 248–260

252

2.5. Discussion

The results of Experiment 1 suggest that when choosing between contrasting statements from the internet or a teacher, second-graders were more likely to endorse the teacher’s statements but kindergarteners displayed no preference. Futhermore, second-graders were more likely to endorse statements from the teacher than kindergarteners. This may be because two years experience offormal education enhanced second-graders’ emotional relationship to their teacher (Deru & Wanlun, 2007), thus, the desire toaffiliate with their teacher may have taken precedence over the positive attributes of the internet. When faced with statementsattributed to the internet or a peer, or statements attributed to a teacher or a peer, neither kindergarterners nor second-gradersconsistently endorsed information from either source. The latter finding is particularly suprising given that second graders endorsedstatements from the teacher more often than those from the internet. One potential explanation for this lack of preference is that thetrivia statements used in Experiment 1 may have seemed similar to the kinds of statements children share with their friends, aschildren can articulate adult-knowledge and child-knowledge and indicate that adults such as teachers are less likely to know theanswer to child-specific knowledge than a child (Fitneva, 2010). Furthermore, VanderBorght and Jaswal (2009) found that althoughchildren acknowledged that other children were likely to be less knowledgeable than adults, there are situations (e.g., questionsabout toys) in which they prefer to seek information from a peer. Experiment 2 was designed to address these issues by only includingscientific and historical facts that involved specific and precise numerical information that children would be unlikely to havediscussed with friends or to think they had personal experience observing.

College students regarded statements from the unspecified internet source and a teacher as more reliable than a peer, yet they didnot have consistent intuitions about which source to endorse when they were presented with contrasting statements attributed to theinternet or a teacher. Due to their long-term experience with formal education, adults may view teachers as possessing more eruditeknowledge (Enesco et al., 2016) than their peers. At the same time, adults’ experience accessing different types of information onlinemay lead adults to consider the internet an equally reliable information source to a trained teacher (Mehrabi et al., 2009).

3. Experiment 2

The trivia statements presented in Experiment 1 represented a wide range of topics. They included statements about culturaldifferences in human behavior (e.g., most Americans like yellow or purple best), which could be considered somewhat subjective orprone to error, and observations about the human body that children might have thought they could answer based on their directexperience or observation (e.g., whether the thumb or the middle finger nail grows the slowest). Due to the nature of the statements,children’s endorsements may have been influenced more by their own intuitions about the correct answer than which source pro-vided the information.1 In order to address this possibility, the questions and statements in Experiment 2 involved only scientific andhistorical facts and focused on specific, numerical values (e.g., how many days it takes a flower to germinate).

In addition, in Experiment 1, participants’ epistemic trust in different informants was operationalized in terms of which in-formant’s statements they endorsed. Although the statements were counterbalanced, participants may have focused more on theirintuitions about the statements than on the information sources, resulting in patterns of responses that did not differ from chance.Thus, in Experiment 2, we also measured participants’ trust in terms of preferences for seeking information from each source beforeasking which statement they endorsed.

3.1. Participants

Thirty-one 5- to 6-year-old kindergarteners (M age = 5;11, range= 5;5–6;2, 16 girls) and 30 7- to 8-year-old second-graders (M age

= 7;7, range=7;0–8;11, 13 girls) participated. Thirty college students (M=20.83 years; SD=2.10; 18 females) also participated.Participants were recruited from the same school and university as Experiment 1. None had participated in Experiment 1.

3.2. Materials

Experiment 2 included 12 scientific and historical questions, which were designed to be unfamiliar to all participants and in-volved numerical values (e.g., How many days does it take Mars to finish a single orbit?; see Appendix B). Each fact was adapted totwo statements that varied in only one numerical value (i.e., “as one of the solar system planets, Mars takes about 600 days to finish asingle orbit” and “as one of the solar system planets, Mars takes about 700 days to finish a single orbit”; see Appendix B for completelist of items). As in Experiment 1, 12 facts were divided into 4 blocks. Each block contained 3 pairs of statements, presented byunspecified internet and a teacher, internet and a peer, a teacher and a peer respectively. The pairings of topics and statements withinformants within each block, and the order in which the blocks appeared was counterbalanced between participants.

3.3. Procedure

The introduction and each informant’s description was exactly the same as in Experiment 1, and children were probed about their

1 Item analyses support that kindergarteners in particular were more likely to endorse the internet for items that involved factual statements thanitems involving more subjective statements (see supplemental Table S2).

F. Wang, et al. Cognitive Development 50 (2019) 248–260

253

online experience as in Experiment 1.

3.3.1. Ask trialsAfter hearing a very brief introduction to each fact statement (e.g., “Look, it is Mars”), participants were first asked whether they

knew the answer of the fact statement (e.g., how many days do you think Mars takes to finish a single orbit?). None of the participantsknew the answer to any of the questions. Then, participants indicated which of two informants they would ask the question (e.g., “Ifyou were going to find out how many days Mars takes to finish a single orbit, which would you like to ask, the internet or theteacher?”; see Appendix B). There were a total of 12 trials. In order to maintain the children’s attention and help them understandeach question, the experimenter also presented a related photograph when the question was introduced.

3.3.2. Endorse trialsFollowing each ask trial, the experimenter presented two contrasting statements about a numerical fact that were attributed to

two of the three informants (e.g., “Teacher says Mars takes about 600 days to finish a single orbit; according to the internet, Marstakes about 700 days to finish a single orbit”; translated from Chinese; see Appendix B). Participants were then asked to endorse oneof the claims (i.e., “Now, how many days do you think Mars takes to finish a single orbit, 600 days or 700 days?”; see Appendix B).The same related photograph was also presented as when the question was introduced.

3.4. Results

3.4.1. Children’s online experience

Following the procedure from Experiment 1, two independent raters coded and classified children’s responses to the onlineexperience questions. There was 100 percent agreement between coders. All children provided appropriate examples of obtaininginformation from the internet, a teacher and a peer. Twenty-eight kindergarteners reported watching videos online (2 additionalchildren’s recordings were incomprehensible or damaged). Sixteen second-graders only reported watching videos online, and 13reported both watching videos and looking up information. As in Experiment 1, no child reported experience using social media.Children’s online experience was also confirmed in interviews with their parents and teachers as in Experiment 1.

3.4.2. Ask trials

Preliminary analyses did not show an effect of gender (Kindergarteners: ts< 1.99, ps> 0.06; Second-graders: ts< 1.81,ps> 0.08) nor item order (Kindergarteners: Fs< 0.61, ps> 0.70; Second-graders: Fs< 2.00, ps> 0.12) on responses on ask trials, sowe did not consider these factors in the rest of the analyses.

3.4.2.1. Internet versus alternate informantsResponses to ask trials were assigned one point for each time the children chose to ask the internet, yielding a score of 0–4 for

each set of trials (internet vs. teacher and internet vs. peer). A 2 (Age: kindergarten, second-grade) × 2 (Alternate Informant: teacher,peer) ANOVA yielded a significant main effect of alternate informant, F(1, 59)= 71.01, p < .001, ηp2 =.546. Bonferroni post-hocanalysis showed that kindergarteners and second-graders were more likely to seek information from the internet when the alternateinformant was a peer (M=3.55) than a teacher (M=1.90). No other main effects nor interactions were significant, F(1, 59) ≤ 2.20,p≥ .144, ηp2 ≤ .036; see Table 2.

Across ask trials where children chose between the internet and the teacher, neither second-graders nor kindergarteners displayeda preference for asking the internet or a teacher (Second-graders: t(29)= 0.76, p= .455, d=0.14; Kindergarteners: t(30)= 0.12,p= .905, d=0.02; see Table 2).

In the trials comparing the internet and a peer, both second-graders and kindergarteners asked the internet significantly moreoften than would be expected by chance (Second-graders: t(29)= 19.20, p < .001, d=3.51; Kindergarteners: t(30)= 5.90,p < .001, d=1.06).

There was no evidence for differences in the likelihood of seeking information from the internet between second-graders whoreported experience looking up information online and second-graders without this experience (see Table S1 in SupplementaryMaterials).

3.4.2.2. Teacher versus peerIn the four trials where the informants were a teacher and a peer, kindergarteners and second-graders indicated that they would

ask the teacher significantly more often than the peer (Kindergarteners: t(30)= 15.58, p < .001, d=2.80; Second-graders: t(29)= 17.90, p < .001, d=3.27; see Table 2). There was no difference between kindergarteners’ and second-graders’ choices, t(59)= 0.83, p= .41, d=0.22.

3.4.2.3. Adults’choicesPreliminary analyses did not show an effect of gender (ts<1.28, ps>0.21) nor item order (Fs<0.76, ps>0.59) on responses on

ask trials, so these factors were not considered in the rest of the analyses. Adults indicated that they would seek information from theinternet at rates significantly more frequent than chance in the internet vs. teacher trials (chance= 2, t(29)= 8.63, p < .001,

F. Wang, et al. Cognitive Development 50 (2019) 248–260

254

d=1.57) and in internet vs. peer trials (t(29)= 26.49, p < .001, d=4.83). When choosing between the teacher and the peer, adultswere more likely to ask the teacher, t(29)= 3.88, p= .001, d=0.71.

3.4.3. Endorse trials

Preliminary analyses did not show effects of gender (Kindergarteners: ts< 1.12, ps> 0.27; Second-graders: ts< 0.97, ps> 0.34)nor item order (Kindergarteners: Fs< 1.24, ps> 0.32; Second-graders: Fs< 2.57, ps> 0.053) on responses in the endorse trials, sothese factors were not considered further.

3.4.3.1. Internet versus alternate informantsEndorse trials were scored as in Experiment 1, yielding a total score of 0–4 indicating how many times the statement from the

unspecified internet source was endorsed on each type of trial (internet vs. teacher and internet vs. peer). A 2 (Age: kindergarten,second-grade) × 2 (Alternate Informant: teacher, peer) ANOVA yielded a significant main effect of alternate informant, F(1,59)= 57.30, p < .001, ηp2 = .493, and an Age×Alternate informant interaction, F(1, 59)= 4.20, p= .045, ηp2 = .066, but nosignificant main effect of age, F(1, 59)= 2.59, p= .113, ηp2 = .042. Bonferroni post-hoc tests revealed that statements from theinternet were more likely to be endorsed when the alternate informant was a peer (M=3.21) than a teacher (M=1.67). Furtheranalysis of the interaction using simple effects tests indicated that both kindergarteners and second-graders were more likely toendorse statements from the internet when the alternate informant was a peer than when it was a teacher (Kindergarten: F(1,59)= 15.50, p < .001; Second-grade: F(1, 59)= 45.50, p < .001).

When choosing between statements from the unspecified internet source and the teacher, second-graders’ and kindergarteners’endorsements did not differ significantly from chance (chance=2; Second-graders: t(29)= 1.49, p= .148, d=0.27;Kindergarteners: t(30)= 1.30, p= .204, d=0.23; see Table 2). In trials involving statements from the internet and the peer, kin-dergarteners and second-graders endorsed the statements from the internet at above-chance levels (Kindergarteners: t(30)= 4.03,p < .001, d=0.73; Second-graders: t(29)= 11.38, p < .001, d=2.08; see Table 2).

As in Experiment 1, there was no evidence that second graders who had experience searching for information online were morelikely to endorse statements from the internet (see Table S1 in Supplementary Materials).

3.4.3.2. Teacher versus peerIn the teacher vs. peer trials, children endorsed the statements made by the teacher at rates well above chance (chance=2;

Kindergarteners: t(30)= 9.32, p < .001, d=1.67; Second-graders: t(29)= 13.38, p < .001, d=2.44). There was no differencebetween kindergarteners and second-graders’ choices, t(59)= 1.09, p= .28, d=0.28.

3.4.3.3. Adults’ choicesPreliminary analyses did not reveal effects of gender (ts< 1.46, ps> 0.16) nor item order (Fs< 2.15, ps> 0.09) on responses in

the endorse trials, so these factors were not considered further. Adults endorsed the answers coming from the internet at ratessignificantly higher than chance in the internet vs. teacher trials (chance= 2, t(29)= 2.97, p= .006, d=0.54) and in internet vs.peer trials (t(29)= 8.57, p < .001, d=1.57). Across four trials where adults chose between the teacher and the peer, they weremore likely to endorse the teacher, t(29)= 7.35, p < .001, d=1.35.

3.4.4. Comparisons between Experiment 1 and Experiment 2

To further investigate whether differences between the nature of the statements used in each experiment impacted children’sjudgments about information from the informants, a 2 (Experiment) × 2 (Age) ANOVA was conducted for the endorse trials for eachcombination of informants.

3.4.4.1. Internet versus teacherIn trials comparing statements from the internet and the teacher, an ANOVA yielded a significant main effect of age, F(2,

117)= 4.58, p= .034, ηp2 = .038. Bonferroni post-hoc analysis showed that second-graders (M=1.52) were less likely to endorsetestimony attributed to the unspecified internet source than kindergarteners (M=1.97). There was no significant main effect ofexperiment, F(1, 117)< 1, p> .10, ηp2< .01, the experiment by age interaction was not significant either, F(1, 117)= 3.17,p= .078, ηp2 = .026.

3.4.4.2. Internet versus peerWhen statements were attributed to the internet or a peer, there was a significant main effect of Experiment, F(1, 117)= 48.29,

p< .001, ηp2 = .292, children were more likely to endorse the internet in Experiment 2 (M=3.22) than in Experiment 1 (M= 1.98).There was also a main effect of age (F(1, 117)= 6.80, p= .010, ηp2 = .055). Post-hoc Bonferroni tests showed that second-graders(M=2.83) were more likely to endorse the internet over a peer than kindergarteners (M=2.37). The interaction was not significant,F(1, 117)= 2.79, p= .097, ηp2 = .023.

3.4.4.3. Teacher versus peerIn the teacher and peer informant trials, there was a significant main effect of Experiment, F(1, 117)= 50.86, p < .001, ηp2 =

F. Wang, et al. Cognitive Development 50 (2019) 248–260

255

.303, but no main effect of age, F(1, 117)= 0.06, p= .815, ηp2< .001. The interaction between Experiment and age was notsignificant, F(1, 117)= 1.96, p= .164, ηp2 = .016. As a whole, children were more likely to endorse the teacher in Experiment 2(M=3.53) than Experiment 1 (M = 2.22).

3.5. Discussion

Overall, 5- to 6-year-old kindergarteners and 7- to 8-year-old second-graders endorsed the teacher’s statements more often thanthe peer’s statements in Experiment 2. This suggests that when faced with judgments about specific numerical values associated withscientific and historical information, children are more likely to realize the limited knowledge of their peers and seek out informationfrom a teacher (see VanderBorght & Jaswal, 2009).

Additionally, in Experiment 2, both kindergarteners and second-graders showed greater trust in unspecified internet source over apeer. This may be because children regard the internet as an information source which contains a wider range of knowledge thantheir peers, a possibility supported by prior work suggesting that children were more likely to trust an adult who provided diverseinformation than one who provided non-diverse information (Landrum, Cloudy et al., 2015; Landrum, Eaves et al., 2015). The resultsalso suggest that children can use the same heuristics that they use to evaluate the reliability of two known individuals (e.g., twoadults) to reason about the credibility of two different types of information sources (e.g., the unspecified internet source and a peer),which provides support for the idea that children’s source evaluation strategies are domain-general.

As for college students, the results in Experiment 2 were inconsistent with the finding in Experiment 1 that adults endorsedstatements from an unspecified internet source and a teacher equally. Confronted with questions and statements involved onlyscientific and historical information, college students were more likely to seek out and endorse information from an unspecifiedinternet source than a teacher. One possible explanation might be that college students believed that although their teachers (i.e.,college professors) could be viewed as a reliable information sources, their knowledge base is not as large as an internet source andtheir memory could not be less reliable or precise than information available on a website. Therefore, when facing questions focusedon specific facts and numerical values, college students were more likely to trust an unspecified internet source.

4. General discussion

The current experiments are among the first to examine children’s preferences for and endorsement of statements from informantsfrom different categories. We explored 5- to 6-year-old kindergarteners’, 7- to 8-year-old second-graders’, and college students’ trustin three informants: an unspecified internet source, a teacher, and a peer. When the statements involved trivia from a broad range ofdomains and included facts that children might think they knew or could directly observe (Experiment 1), kindergarteners did notshow consistent endorsement of statements from any of the informants and second-graders only showed a preference for endorsingstatements from a teacher over those from the internet. When the statements were about objective, numerical values associated withscientific and historical information (Experiment 2), both kindergarteners and second-graders were more likely to say that they wouldseek out answers from the internet or a teacher than from a peer, and their endorsements showed the same pattern. In contrast, adultsshowed greater trust in the internet or a teacher than a peer across both experiments, and they also trusted the unspecified internetsource more than a teacher when evaluating statements about scientific and historical facts in Experiment 2.

Taken together, our findings suggest that school age children can reason about the reliability of information sources from differentcategories and that their judgments are sensitive to the type of information being sought. Just as children can be skeptical whenmaking judgments about the reliability of different people (see Mills, 2013), children’s belief in information from the internet is notimmutable. Rather, children’s assessment of the trustworthiness of an internet-based source may depend on the circumstances. Forinstance, when the information in question involves specific numerical values associated with scientific phenomena that childrencannot directly observe, children are more likely to rely on the internet, but when the information is from other domains (e.g.,information that seems more accessible or that can be directly observed), they do not view the unspecified internet as more trust-worthy than a peer. Children may also consider facts about specific quantities or dates, such as those in Experiment 2, to be morereliably addressed by the internet that facts about behaviors and appearances, such as those in Experiment 1. That said, additionalresearch is needed to examine how the nature and domain of the information at hand, and children’s prior familiarity with theinformation, may play into children’s evaluations of the reliability of different informants (see also Danovitch & Keil, 2007).

Prior work on the development of epistemic trust suggests that children use at least two distinct strategies when selecting betweeninformants who provide conflicting information. The first is to focus on their epistemic goals by making judgments based on theinformants’ previous accuracy or the extent of their knowledge (see Harris, 2012 for a review). The second is to focus on socialidentity which can help children establish and maintain positive social relationships. In these cases, children show greater trust ininformants who are familiar, or who share features, such as an accent or group membership, with them (e.g., Corriveau & Harris,2009; Corriveau, Kinzler, & Harris, 2013). In the current experiments, when children chose between endorsing statements from theinternet or a peer, children’s preference for the internet (in Experiment 2) suggests that they put more weight on epistemic char-acteristics than social ones. However, across both experiments, children showed similar levels of endorsement of statements from theinternet and a teacher (as did adults in Experiment 1), which may reflect a view of both informants as similarly capable of answeringquestions. This finding might also suggest that participants’ socially driven desire to treat the teacher as an intellectual authoritycompensated for the teacher’s potential epistemic limitations in comparison to the internet source.

In addition, our findings indicate that there are major differences in how children and adults treat information from the internet.One possible explanation for this age difference may be a function of increasing exposure to the diverse types of information available

F. Wang, et al. Cognitive Development 50 (2019) 248–260

256

on the internet and a growing appreciation of the epistemic capacity of the internet. In the current study, information from theinternet was described as information from online video and news websites (including information from a popular search engine) tofit children’s limited practical online experience. Adults are more likely to realize the complexity of the internet (Yan, 2005, 2006)and understand the value of the internet in terms of access to much more information than any person could know. Likewise, theremay also be less variability in how adults conceptualize the internet and its capabilities. Another possible explanation might be thatas children mature into adults, they gain a greater understanding of deception, and realize that intentions and motivations couldinfluence a human informant’s reliability (see Mills, 2013 for a review). Thus, adults show less trust in other humans, even if they areteachers. Finally, differences in adults’ and children’s judgments may also depend on differing views of the consequences of endorsingthe wrong answer. Choosing the wrong answer to the questions in our experiments may not have seemed very consequential forchildren, but adults may have more experience with situations where endorsing incorrect information may be costly (e.g., in a collegecourse or a job setting). Thus, adults may place more emphasis on a source’s likely accuracy than on maintaining social bonds orrespecting authority.

Our findings suggest that Chinese children may experience a tension at times between trusting a familiar adult or the internet2,which could have implications for their learning. Although all of the children in our studies had experience with the internet,relatively few reported having searched for information online. Given that even college students struggle to evaluate informationfound online (Mason, Boldrin, & Ariasi, 2009), elementary school children may benefit from explicit instruction about the nature ofthe internet and from teacher-guided examples of how to use the internet to obtain information. One of the practical implications ofour findings is that young children may be reluctant to choose alternate types of information sources over a familiar, trustworthyadult, and that it may be necessary to support them or explicitly instruct them that there are benefits to seeking out information fromdifferent types of sources. For example, watching a teacher or a parent model an online information search and apply appropriatecriteria to evaluating the results (e.g., looking for the author of a website; see Mason, Junyent, & Tornatora, 2014) could be aneffective way of introducing children to the concept that the internet can provide information that even an knowledgeable adultwould not know.

Since the internet is unlike all other information sources in terms of the quantity and speed at which information can be accessed(Gerjets & Hellenthal-Schorr, 2008), and children are not adept at evaluating the reliability of online information (Kuiper, Volman, &Terwel, 2008), it is important for educators and caregivers to translate and filter online information for their children (especially foryounger children). However, in the present experiments, when children’s parents were interviewed, none of them indicated that theywould always accompany their children while the children viewed online videos via PC or other portable electronic devices. In mostcases, Chinese children just go online by themselves. Given that instructional interventions can be effective at improving children’sevaluation behavior (Mason et al., 2014), it would be interesting to examine whether encouraging parent’s filtering or translation ofinformation from internet sources influences children’s trust in the internet in the future.

One potential limitation of the current studies is that the statements in question involved trivia or scientific and historical factswhich were designed to be novel to the Chinese children who participated. Although related photographs were used to help childrenvisualize and understand these statements, we did not directly measure children’s comprehension of the statements. Recent evidencesupports that, compared to children already had some background knowledge about a target question (though they still did not knowthe answer), 5- to 7-year-old children who were naïve about a target question were more likely to ask a teacher who could provideexhaustive information (Gweon, Shafto, & Schulz, 2018). Similarly, it is possible that, in our studies, children who lacked backgroundknowledge about the statement topic or who could not fully comprehend the statements were more likely to seek out answers fromthe internet for its abundance of information (in the ask trials of Experiment 2). However, had the statements involved more familiartopics, children may have been more likely to trust individuals who could efficiently provide an accurate answer (i.e., a teacher) thaninformation sources that could access a broader range of knowledge. Consequently, future research might investigate whetherchildren’s familiarity with and comprehension of the topics in question relate to the strategy they use to determine which informationsource they ask or endorse.

Another potential limitation is that unlike children’s limited online experiences, adults’ forms of internet use were much morediverse. Therefore, college student participants may have viewed the unspecified internet source as potentially including a vastcollection of websites or videos while they considered a teacher to be a specific known individual (i.e., their college professor). Giventhe great variability in the credibility of information found online, future studies might only focus on college students and investigatetheir epistemic trust in a teacher versus one specific kind of internet source (e.g., information published by governmental institutions,or advertisements from a commercial site).

Finally, it is important to note that the participants in this study were Chinese children. Given that “honoring the teacher” andrespecting authority is deeply emphasized in Chinese socio-cultural traditions, children’s trust in the teacher may have reflectedsocio-cultural practices and expectations rather than a genuine belief that the teacher was more likely to be accurate. It would beinformative for future research to compare trust in teachers or internet-based sources among Chinese children and children in othercultures where questioning the teacher’s and adults’ authority is a more socially acceptable practice. In Western culture, associatedwith Socratic ideals for learning, children are encouraged to question their own and others’ beliefs and evaluate others’ knowledge(see Tweed & Lehman, 2002 for a review). Parents of Western children also tend to raise more why-questions in the interactions with

2 In Experiment 2, we calculated the number of inconsistencies in responses between the ask and endorse trials. These inconsistencies were mostprevalent in the internet vs. teacher trials, suggesting that participants were uncertain about which informant’s response to endorse even afterindicating a preference to ask one of the informants (see supplemental Table S3).

F. Wang, et al. Cognitive Development 50 (2019) 248–260

257

their children (Gauvain, Munroe, & Beebe, 2013), which might be more conducive to the development of argumentation skill.Consequently, Western children may be more likely to realize the limited knowledge of human informants and a Western sample mayput more weight on epistemic factors and show greater trust in the internet than a teacher or a peer.

The results of the current study suggest that school-aged children exhibit flexible trust in an unspecified internet source, teacher,and peer. Moreover, trust in statements from the internet may be affected by the type of information being evaluated, such that anunspecified internet source is considered more credible in some domains (e.g., scientific and historical facts) and for certain types ofinformation (e.g., numerical values) than others. As an initial foray into the development of children’s judgments about the reliabilityof information from the internet and known individuals, these findings offer an insight into the factors that influence children’s trustin the internet, and provide a foundation for future research focusing on how children’s experience with and understanding of novelinformation sources, such as the internet, influences their preferences and judgments.

Funding

This research was supported by the National Natural Science Foundation of China Grant (#31771236, 31700968) andFundamental Research Funds for the Central Universities(#CCNU19TS040).

Appendix A

Table A1

Appendix B

Table B1

Table A1Twelve questions used in Experiment 1 (Translated from Chinese).

Block Statement 1 Statement 2 Test

1 Most Americans like yellow best. Most Americans like purple best. What color do you think American like best, yellow orpurple?

Thumb nail grows the slowest. The middle finger nail grows the slowest. Which finger nail do you think grows the slowest, thenumb or the middle finger?

All polar bears like using their right hand. All polar bears like using their left hand. Which hand do you think polar bears like to use, right orleft?

2 American use toothpicks most often. German use toothpicks most often. Who use toothpicks most often, American or German?The color of lobster's blood is blue. The color of lobster's blood is green. What color do you think lobster’s blood is, blue or green?When eating bananas, people are morelikely to be bitten by mosquitoes.

When eating apples, people are more likelyto be bitten by mosquitoes.

Which fruit makes people more likely to be bitten bymosquitoes, banana or apple?

3 The cat has more fingers on its back paws. The cat has more fingers on its front paws. Which paws do cat has more fingers on, back paws orfront paws?

Magpie is the only kind of bird which candistinguish blue from other colors.

Owl is the only kind of bird which candistinguish blue from other colors.

Which kind of birds can distinguish blue from other color,magpie or owl?

Girl’s hearts beats faster. Boy’s hearts beats faster. Whose heart beats faster, girl or boy?4 Information is better remembered when it

was heard by the right ear.Information is better remembered when itwas heard by the left ear.

Through which ear do you think the information heard isbetter remembered, the right ear or the left ear?

The first bottle of cola in the world is blue. The first bottle of cola in the world isgreen.

What color do you think the first bottle of cola is, blue orgreen?

The earth moves around the sun faster inwinter.

The earth moves around the sun faster insummer.

Which season does earth move around the sun faster,winter or summer?

Table B1Twelve scientific and historical facts used in Experiment 2 (Translated from Chinese).

Block Ask Trials Endorse Trials

1 If you are going to find out how many days Mars takes to finish asingle orbit, who would you like to ask, ×× or ××?

×× says Mars takes about 600 days to finish a single orbit; ×× says Marstakes about 700 days to finish a single orbit. Now, how many days do you thinkMars takes to finish a single orbit, 600 days or 700 days?

If you are going to find out how many layers in the earth, who wouldyou like to ask, ×× or ××?

×× says the earth is constituted with 4 layers; ×× says, the earth isconstituted with 5 layers. Now, how many layers do you think earth isconstituted with, 4 layers or 5 layers?

If you are going to find out how many years mammoth have existed,who would you like to ask, ×× or ××?

×× says mammoth have existed for 4.5 million years; ×× says mammothhave existed for 4.6 million years. Now, how many years do you thinkmammoth have existed for, 4.5 million years or 4.6 million years?

(continued on next page)

F. Wang, et al. Cognitive Development 50 (2019) 248–260

258

Appendix C. Supplementary data

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.cogdev.2019.05.006.

References

Bascandziev, I., & Harris, P. L. (2014). In beauty we trust: Children prefer information from more attractive informants. British Journal of Developmental Psychology,32(1), 94–99.

Bascandziev, I., & Harris, P. L. (2016). The beautiful and the accurate: Are children’s selective trust decisions biased? Journal of Experimental Child Psychology, 152,92–105.

Bergin, C., & Bergin, D. (2009). Attachment in the classroom. Educational Psychology Review, 21(2), 141–170.Brosseau-Liard, P. (2017). The roots of critical thinking: Selective learning strategies in childhood and their implications. Canadian Psychology / Psychologie canadienne,

58(3), 263–270.China Internet Network Information Center (2018a). The 42nd China statistical report on Internet developmentRetrieved fromhttp://www.cnnic.cn/hlwfzyj/hlwxzbg/

hlwtjbg/201808/t20180820_70488.htm.China Internet Network Information Center (2018b). Whitepaper: China internet statistics. Retrieved fromhttps://www.chinainternetwatch.com/whitepaper/china-

internet-statistics/.China National Youth Palace Association (2015). Research report on internet use of Chinese teenagersRetrieved fromhttp://www.cnypa.org/dcyj/983828.jhtml.Corriveau, K. H., & Harris, P. L. (2009). Choosing your informant: Weighing familiarity and recent accuracy. Developmental Science, 12(3), 426–437.Corriveau, K. H., Einav, S., Robinson, E. J., & Harris, P. L. (2014). To the letter: Early readers trust print-based over oral instructions to guide their actions. British

Journal of Developmental Psychology, 32(3), 345–358.Corriveau, K. H., Kinzler, K. D., & Harris, P. L. (2013). Accuracy trumps accent in children’s endorsement of object labels. Developmental Psychology, 49(3), 470–479.Danovitch, J. H., & Alzahabi, R. (2013). Children show selective trust in technological informants. Journal of Cognition and Development, 14(3), 499–513.Danovitch, J. H., & Keil, F. C. (2007). Choosing between hearts and minds: Children’s understanding of moral advisors. Cognitive Development, 22(1), 110–123.Danovitch, J. H., & Mills, C. M. (2014). How familiar characters influence children’s judgments about information and products. Journal of Experimental Child

Psychology, 128, 1–20.Deru, S., & Wanlun, L. (2007). A study of the characteristics of the development of teacher-student relations in middle and primary schools. Journal of Psychological

Science, 30(4), 873–877.Dodge, A. M., Husain, N., & Duke, N. K. (2011). Connected kids? K-2 children’s use and understanding of the internet. Language Arts, 89(2), 86–98.Doebel, S., Rowell, S. F., & Koenig, M. A. (2016). Young children detect and avoid logically inconsistent sources: The importance of communicative context and

executive function. Child Development, 87(6), 1956–1970.Duarte Torres, S., Weber, I., & Hiemstra, D. (2014). Analysis of search and browsing behavior of young users on the web. ACM Transactions on the Web, 8(2), 1–54.Eastin, M. S., Yang, M.-S., & Nathanson, A. I. (2006). Children of the Net: An empirical exploration into the evaluation of internet content. Journal of Broadcasting &

Electronic Media, 50(2), 211–230.Einav, S., Robinson, E. J., & Fox, A. (2013). Take it as read: Origins of trust in knowledge gained from print. Journal of Experimental Child Psychology, 114(2), 262–274.Elashi, F. B., & Mills, C. M. (2014). Do children trust based on group membership or prior accuracy? The role of novel group membership in children’s trust decisions.

Journal of Experimental Child Psychology, 128, 88–104.Enesco, I., Rodríguez, P., Lago, M. O., Dopico, C., & Escudero, A. (2016). Do teachers’ conflicting testimonies influence children’s decisions about unconventional rules

Table B1 (continued)

Block Ask Trials Endorse Trials

2 If you are going to find out how many teeth each person has, whowould you like to ask, ×× or ××?

×× says each person has 34 teeth; ×× says each person has 36 teeth. Now,how many teeth do you think each person has, 34 teeth or 36 teeth?

If you are going to find out what their body temperature is, whowould you like to ask, ×× or ××?

×× says bird’s temperature is 40 degrees Celsius; ×× says bird’s temperatureis 42 degrees Celsius. Now, what do you think bird’s temperature is, 40 degreesCelsius or 42 degrees Celsius?

If you are going to find out how many days the seeds of sunflowerstake to germinate, who would you like to ask, ×× or ××?

×× says the seeds of sunflowers take about 8 days to germinate; ×× says theseeds of sunflowers take about 10 days to germinate. Now, how many days doyou think the seeds of sunflowers take to germinate, about 8 days or 10 days?

3 If you are going to find out how many bones your hands consists of,who would you like to ask, ×× or ××?

×× says one hand is composed of 26 thin bones; ×× says one hand iscomposed of 27 thin bones. Now, how many thin bones do you think your handconsists of, 26 thin bones or 27 thin bones?

If you are going to find out how many years the origin of India can betraced back to, who would you like to ask, ×× or ××?

×× says the origin of India can be traced back to 4200 years; ×× says theorigin of India can be traced back to 4500 years. Now, how many years do youthink the origin of India can be traced back to, 4200 years or 4500 years?

If you are going to find out the exact date when the contour of themoon is like this image, who would you like to ask, ×× or ××?

×× says the shape of the lunar in Jan. 2nd is like this image; ×× says theshape of the lunar in Jan. 28th is like this image. Now, in which date do youthink the contour of the moon is like this image, Jan. 2nd or Jan. 28th?

4 If you are going to find out how large Australia is compared to theother countries in the world, who would you like to ask, ×× or××?

×× says Australia is the 6th largest country in the world; ×× says Australia isthe 7th largest country in the world. Now, how large do you think Australia iscompared to the other countries, the 6th or the 7th?

If you are going to find out how many muscles each person has, whowould you like to ask, ×× or ××?

×× says each person has around 500 muscles; ×× says each person hasaround 600 muscles. Now, how many muscles do you think each person has,about 500 muscles or about 600 muscles?

If you are going to find out how long the plane has been invented,who would you like to ask, ×× or ××?

×× says the plane has been invented for 110 years; ×× says the plane hasbeen invented for 120 years. Now, how long do you think the plane has beeninvented, for 110 years or for 120 years?

F. Wang, et al. Cognitive Development 50 (2019) 248–260

259

of counting? European Journal of Psychology of Education, 1–18.Eyden, J., Robinson, E. J., & Einav, S. (2014). Children’s trust in unexpected oral versus printed suggestions: Limitations of the power of print. British Journal of

Developmental Psychology, 32(4), 430–439.Eyden, J., Robinson, E. J., Einav, S., & Jaswal, V. K. (2013). The power of print: Children’s trust in unexpected printed suggestions. Journal of Experimental Child

Psychology, 116(3), 593–608.Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences.

Behavior Research Methods, 39(2), 175–191.Fitneva, S. A. (2010). Children’s representation of child and adult knowledge. Journal of Cognition and Development, 11(4), 458–484.Flanagin, A. J., & Metzger, M. J. (2000). Perceptions of internet information credibility. Journalism & Mass Communication Quarterly, 77(3), 515–540.Gauvain, M., Munroe, R. L., & Beebe, H. (2013). Children’s questions in cross-cultural perspective: A four-culture study. Journal of Cross-Cultural Psychology, 44(7),

1148–1165.Gerjets, P., & Hellenthal-Schorr, T. (2008). Competent information search in the World Wide Web: Development and evaluation of a web training for pupils. Computers

in Human Behavior, 24(3), 693–715.Giles, J. (2005). Internet encyclopaedias go head to head. Nature, 438, 900–901.Gweon, H., Shafto, P., & Schulz, L. (2018). Development of children’s sensitivity to overinformativeness in learning and teaching. Developmental Psychology, 54(11),

2113–2125.Harris, P. L. (2012). Trusting what you’re told: How children learn from others. Cambridge, MA: The Belknap Press/Harvard University Press.Harris, P. L., Koenig, M. A., Corriveau, K. H., & Jaswal, V. K. (2018). Cognitive foundations of learning from testimony. Annual Review of Psychology, 69(1), 251–273.Hart, P. E., & Liu, Z. (2003). Trust in the preservation of digital information. Communications of the ACM, 46(6), 93–97.International Society for Technology in Education (1998). National education technology standards for students. Eugene, OR: International Society for Technology in

Education.International Society for Technology in Education (2007). National education technology standards for students (2nd ed.). Eugene, OR: International Society for

Technology in Education.Jaswal, V. K., & Neely, L. A. (2006). Adults don’t always know best: Preschoolers use past reliability over age when learning new words. Psychological Science, 17(9),

757–758.Johnson, M. K. (2007). Reality monitoring and the media. Applied Cognitive Psychology, 21(8), 981–993.Johnston, A. M., Mills, C. M., & Landrum, A. R. (2015). How do children weigh competence and benevolence when deciding whom to trust? Cognition, 144, 76–90.Koenig, M. A., & Harris, P. L. (2005). Preschoolers mistrust ignorant and inaccurate speakers. Child Development, 76(6), 1261–1277.Koenig, M. A., Clément, F., & Harris, P. L. (2004). Trust in testimony: Children’s use of true and false statements. Psychological Science, 15(10), 694–698.Koomen, H. M., & Hoeksma, J. B. (2003). Regulation of emotional security by children after entry to special and regular kindergarten classes. Psychological Reports,

93(3 Pt. 2), 1319.Kuiper, E., Volman, M., & Terwel, J. (2008). Integrating critical web skills and content knowledge: Development and evaluation of a 5th grade educational program.

Computers in Human Behavior, 24(3), 666–692.Landrum, A. R., Cloudy, J., & Shafto, P. (2015). More than true: Developmental changes in use of the inductive strength for selective trust. Conference of the Cognitive Science

Society.Landrum, A. R., Eaves, B. S. J., & Shafto, P. (2015). Learning to trust and trusting to learn A theoretical framework. Trends in Cognitive Sciences, 19(3), 109–111.Li, Q.-G., Heyman, G. D., Xu, F., & Lee, K. (2014). Young children’s use of honesty as a basis for selective trust. Journal of Experimental Child Psychology, 117, 59–72.Mantzicopoulos, P., & Neuharth-Pritchett, S. (2003). Development and validation of a measure to assess head start children’s appraisals of teacher support. Journal of

School Psychology, 41(6), 431–451.Mason, L., Boldrin, A., & Ariasi, N. (2009). Searching the web to learn about a controversial topic: Are students epistemically active? Instructional Science, 38(6),

607–633.Mason, L., Junyent, A. A., & Tornatora, M. C. (2014). Epistemic evaluation and comprehension of web-source information on controversial science-related topics:

Effects of a short-term instructional intervention. Computers & Education, 76, 143–157.Mehrabi, D., Hassan, M. A., & Ali, M. S. S. (2009). News media credibility of the internet and television. European Journal of Social Sciences, 11(1), 136–148.Mertala, P. (2019). Young children’s conceptions of computers, code, and the Internet. International Journal of Child-computer Interaction, 19, 56–66.Metzger, M. J., & Flanagin, A. J. (2013). Credibility and trust of information in online environments: The use of cognitive heuristics. Journal of Pragmatics, 59, 210–220.Metzger, M. J., Flanagin, A. J., Markov, A., Grossman, R., & Bulger, M. (2015). Believing the unbelievable: Understanding young people’s information literacy beliefs

and practices in the United States. Journal of Children and Media, 9(3), 325–348.Mills, C. M. (2013). Knowing when to doubt: Developing a critical stance when learning from others. Developmental Psychology, 49(3), 404–418.More than 60% Chinese children have mobile phones, leading parents’ and educators’ worry about network security. Retrieved from Guangming Daily http://mini.eastday.

com/a/171204102204450-3.html.More than 90% of children get access to internet from the age of 3. Retrieved from http://baby.163.com/16/0926/12/C1SV5VUV00367V72.html.Palmquist, C. M., Jaswal, V. K., & Rutherford, A. (2016). Success inhibits preschoolers’ ability to establish selective trust. Journal of Experimental Child Psychology, 152,

192–204.Robinson, E. J., Einav, S., & Fox, A. (2013). Reading to learn: Prereaders’ and early readers’ trust in text as a source of knowledge. Developmental Psychology, 49(3),

505–513.She, Y., Liu, F., & Fan, Z. (2018). A survey of developments on the characteristics of the Internet information. Journal of Nanjing University of Information Science &

Technology, 10(4), 409–414.St. Jean, B., Taylor, N. G., Kodama, C., & Subramaniam, M. (2017). Assessing the digital health literacy skills of tween participants in a school-library-based after-

school program. Journal of Consumer Health on the Internet, 21(1), 40–61.Tweed, R. G., & Lehman, D. R. (2002). Learning considered within a cultural context: Confucian and Socratic approaches. American Psychologist, 57(2), 89–99.van Laar, E., van Deursen, A. J. A. M., van Dijk, J. A. G. M., & de Haan, J. (2017). The relation between 21st-century skills and digital skills: A systematic literature

review. Computers in Human Behavior, 72, 577–588.VanderBorght, M., & Jaswal, V. K. (2009). Who knows best? Preschoolers sometimes prefer child informants over adult informants. Infant and Child Development, 18(1),

61–71.Waldrop, M. M. (2017). News feature: The genuine problem of fake news. Proceedings of the National Academy of Sciences, 114, 12631–12634.Yan, Z. (2005). Age differences in children’s understanding of the complexity of the Internet. Applied Developmental Psychology, 26(4), 385–396.Yan, Z. (2006). What influences children’s and adolescents’ understanding of the complexity of the Internet? Developmental Psychology, 42(3), 418–428.

F. Wang, et al. Cognitive Development 50 (2019) 248–260

260